dive (was Re: Sos-)

I was planning to leave this trainwreck for now, but what the heck, this seems to be getting into some key issues, so here goes.

> > My factual objection IS that you in fact have provided no factual
> > argument.
>
> That is not a factual objection to any of the assumptions I make, but an unspecified objection to all of them.

Yes, it is an objection to your general methodology. Do you have a problem with me doing that?

> http://en.wikipedia.org/wiki/Karl_Popper#Problem_of_Induction
> '...no assumption can ever be or needs ever to be justified, so a lack of justification is not a justification for doubt'

And thus, these days by "justification" we should not understand "proof", as they still apparently did in Popper's time, but "demonstration that the assumption economically explains stuff".

> > Re: your reply to Brian, let me clarify I'm not requesting
> > "proof" and at no point have I;
>
> Why talk about 'burden of proof' then?

Since it's accepted standard terminology.
http://en.wikipedia.org/wiki/Philosophic_burden_of_proof

> You got some nerve - first you weasel switch the definition of what you said earlier

No, my definition has remained constant, but I'm glad to see your understanding of it has progressed beyond literalism.

> > translated to your Popperist terminology, this is indeed
> > essentially a request to "show that your assumption explains more
> > with less than the existing theories".
>
> Ok, so now you redefine what you meant by 'burden of proof'.

Not "redefine", I'm transcribing my point to your terms since you seem to be incapable of talking in any others.

> > you apparently only explain more data by using MORE assumptions
> > than the theories thus far, which is a non-achievement.
>
> Okay, so this is now your objection to my *presentation* of my theory.

If it's only a presentation issue, you really need to work on that.

(Tho it IS nothing new that you rarely bother to explicitely state much about your thesis, beyond "I think this huge list of words is somehow related")

> to defeat a theory, you can do three things:
>
> 1) show that it doesn't show what it purports to show, or that there is logical flaw in its inferences (primary method)
> 2) show that some other theory shows more with the same (secondary)
> 3) show that some other theory shows the same with less (secondary)

I agree loosely, tho I would call scenario 2) an extension rather than a defeat of the theory, and split 1) into its two constituent cases. Of these the first is the one I think you're failing the hardest on, tho depending on what exactly it is you purport to have shown.

> The competition my theory has is two:
>
> 1) The theory that there were no substrate languages in NW Europe or if there were they left no marks on either IE or Uralic, therefore only inner developments within IE and Uralic should be posited for any word in those languages.
>
> 2) The theory that there existed besides PIE and Uralic in NW Europe also
> i) Kuhn's ur-/ur- language (NWBlock I),
> ii) Schrijver's language of geminates, and
> iii) Schrijver's language of bird names

False trichotomy. Numerous other explanations are possible, for example that Kuhn's ar/ur-language did exist, but the "language of geminates" is simply an artifact of Germanic/Celtic/other languages of the area loaning from substrates that didn't contrast gemination and representing their consonants as sometimes singletons, sometimes geminates. Remember, that some theories have not been proposed yet does not mean that no further theories would be possible.

> As for theory 1 I don't consider it competition; there is sufficient consensus for me that there must have been now extinct languages in NW Europe influencing PIE and Uralic.

Yep.

> As for theory 2 I have proposed that i - iii are one and the same language, since the vocabularies ascribed to them overlap. So my theory explains the same with less than 2)

Less explicitely assumed languages, but more assumptions of phonetic developments. And as there most likely were numerous substrate languages anyway, getting the number of languages down is not a particularly pressing concern.

Actually, you've also told you do not consider there having been a single language, but several related languages, so you don't even reach that. So what does your theory exactly have less of?

I don't see vocabulary overlap as much of problem as there is no law that would prohibit substrates having loaned words from one another.

Anyway, the existence of N specific substrates is not all you're up against. In most cases (eg. *saiwa vs. *sal) I think you've not even managed to defeat the null hypothesis of "unrelated".

> > > > > The problem is, the semantic dividing lines you can set up
> > > > > for my proposed set of reflexes of *saN-, do not match
> > > > > similar semantic dividing lines in reflexes of *daN-.
> > > >
> > > > Semantic dividing lines are not required to conform to any
> > > > pattern if these are originally unrelated roots.
> > >
> > > Yes, that is your assumption, but you can't use your assumption
> > > to prove things, as you pointed out yourself.
> >
> > True, but there's nothing I'm trying to proov.
>
> Well, you're trying to disprove.

No. Disproof doesn't really exist, so far as "proof" can only be understood as "demonstration". I'm batting for the null hypothesis here.

If it makes you happier, I can concede that everything you've proposed is *possible*. As is that you're about to be lick'd by an invisible dragon. But as a skeptic I don't really care about mere possibility, only plausibility.

> > The "semantic dividing lines do not match"? So what?
>
> So your argument doesn't work

I do not follow. AFAIK I've based no argument on "semantic division lines matching".

> This

(…alternations…)

> is how the phonetic rule in
> http://tech.groups.yahoo.com/group/cybalist/message/65884
> should have been presented, it is not dependent the preceding *d-

Yeah, I kno.

> > > An example is the dividing line you set up between the "suck"
> > > and the "sump" roots.
> >
>
> That was an example of a semantic division set up by you, the one between what might be called the "suck" words and the "sump" words.

It is not set up by me: "suck" continues to not mean the same as "swamp", but I guess if you dislike that, you do have the right to ignore reality.

> de Vries
> 'súgr m. 'sea' (poet.), lit. 'the sucking (one)'

"Sea" does not equate with "swamp". "Swamp" does not equate with "sucking". "Sucking" does not equate with "to suck".

> > And what is this similar to, and in what way?
>
> You can't set up a corresponding dividing line between words meaning "duck" and "dump",

Why should there be? I've not argued that the structure of the "*daN" and "*saN" sets is the same.

> > > > > If you are convinced there's nothing to be solved, what are
> > > > > you doing in linguistics?
> > > >
> > > > There's a distinction between "not convinced that X" (aka
> > > > skepticism) and "convinced that not X" (aka disbelief).
> > >
> > > And how is this relevant?
> >
> > In that I'm the latter, not former as you just implied.
>
> You don't mean the other way around?

Nope. You're not confused on which of "former" and "latter" is which?

> > > > > They have to with "soul", "immortality", "truth" and "the
> > > > > otherworld". That's why it's interesting to find out how
> > > > > they are related.
> > > >
> > > > "Salt", "saliva", "island", "slush" have nothing to do with
> > > > those topics.
> > >
> > > The guy who first proposed a connection between *saiwa- "lake"
> > > and *saiwala- "soul" is the one who committed the original sin.
> > > I'm just trying to find a semantic connection between them.
> >
> > According to what you've cited so far that one means not simply
> > "lake", but "sacred lake", "water used for divination". This DOES
> > have to do with "soul", "truth" etc.
>
> Yes, and therefore I also believe those semantic fields I mentioned are connected.

All semantic FIELDS are "connected". The question is how much semantic ground some specific WORD covers. In this case, *saiwa covers both hydronymic and sacred territory; *sal- "island" only covers the former. As far as hydronyms go, it also only covers "island", and *saiwa only covers "lake". No overlap.

You seem to have skipped over the question of IF these are related at all, before venturing into the "how". Sure, you can just assume a relationship, but your results can only be as strong as your weakest assumption.

Then again, I'm still not sure if you've claimed any actual results.

> > > The rest are trivial.
> >
> > They are not. Connecting *saiwa and *saiwala requires no
> > assumptions about sound changes affecting the root, only
> > suffixation of /-la/; likewise we require only minor semantical
> > changes ("soul" probably suffices as an original to begin with).
> > Connecting something like "salt" by contrast requires numerous
> > assumptions of phonetical and semantical changes.
>
> You don't need to postulate anything drastic semantically to connect "salt" and "sacred"
> http://en.wikipedia.org/wiki/Salt#In_religion

I said "numerous", not "drastic". Certainly the assumption of a religion in which salt plays a role is possible, as is that speakers of such a culture could have etymologically connected words for "sacred" and "salt". This still doesn't stop these from BEING new requisite assumptions. You do see that, do you?

"Requiring assumptions not previously made" would make a good definition of "non-trivial" for me.

> > And... you just outlined how class VI verbs would in your view
> > have develop'd within Germanic; ie. the a/u alternation does not
> > manifest itself as a/o: after all. Now what?
>
> So the a/o: alternation must have been present in PPGmc as a/a:, which is not an IE thing, which indicates the words involved were loans.

I've not been arguing against loan origin, I've been arguing against your previous claim that the a/o: alternation is in some fashion the same as (or as you put it, "manifested as") the a/u alternation.

> > > > > Verbs loaned at that time would have been too drastically
> > > > > changed to be recognized if patterned on the existing
> > > > > ablauting verbs,
> >
> > > > There's no reason the loan-givers would still have to
> > > > recognize the word,
> > >
> > > ? No, but the loan-receivers would.
> >
> > Yeah, so they recognize it, and proceed to inflect it in an usual
> > sense. If they can uphold ablaut as productiv, they can recognize
> > each form involved in it.
>
> That might be the case for the triliteralized loans in Hebrew you mentioned. It's not a universal law.

Just about by definition, productiv alternations are recognizable. What exactly you're trying to say? That Proto-Germans inflected these words as class VI because they didn't understand their own forms in other classes anymore?

> > > Ok, so you argue that the source of the a/u alternation might
> > > be a non-nasalized vowel *o, and now you nasalize it again?
> >
> > I am proposing eg. the existence of both non-nasal *o, and nasal
> > *o~ (as a possibility to explain ). Or equivalently, a single *o
> > and a biphonemic structure such as *on.
>
> Occam. You explain the same with more.
>
> You need to assume the existence of two vowels instead of one. Fail.

Ah, I see, you're ignoring the principle of uniformitarianism. EVERY human language that has nasal vowels, also has corresponding oral vowels. So your supposed substrate must've had too - or otherwise you're covertly assuming it to have been fundamentally different from all known human languages. This is a considerably bigger assumption than the existence of an oral vowel phone(me).

> > > These words have variants with nasals.
> >
> > What "these words"? I am precisely speaking of cases where there
> > AREN'T variants with nasals, e.g. "suck".
>
> Da. synke (< *sinkW- < *senkW- ?< *sunkW- < *sunk-; sank, sunket class III, but cf Sw. sjunka, sjönk, sjunkit, class II, recategorized or ?) means
> 1. "to sink" (intr.)
> 2. "to swallow" (tr. and intr.)

"Sink, swallow" does not equate with "suck". You have difficulties respecting semantic boundaries, that's for certain.

> Sogn m. ON. 'Sognefjord', auch 'see' (poet.),
> vgl. schw. sugn 'Wasserfall'.
> â€" Nach der saugung des wassers genannt
> (s O. Rygh, Norske Gaardnavne 1, 325),
> zu súga.

A derivative "sucking", not the SAME word as "suck".

> lit. sunkiù, súñkti 'absickern lassen',
> lett. su:zu, su:kt 'saugen'.

I don't kno enuff about Baltic to see if this adds up to a single Proto-(East) Baltic root, or if we have to set up two distinct roots (with the latter then belonging together with Germanic "suck", and the former left hanging or related to "sink").

(...)

> The irregularity of the vowel of the first syllable in Saami is not without parallels; a *u in the protolanguage many times corresponds also to Saami â,
>
> Note: a/u alternation badly explained.

> Torsten

Northern Sami â comes from Proto-Samic *ë (an unrounded mid back vowel, remains as such in Eastern Sami) which regularly comes from PU *i. PU *u regularly becomes Proto-Samic *o. Confirmation bias at work: at no point is there an a/u alternation here.

This particular alternation is one of mid back vowel labiality, ie. between PSm *ë and *o. It occurs mostly in the vicinity of labial consonants. Seems about as well-explained as an irregular vowel alternation can be.

John Vertical